Detecting Fraudulent Transactions in Automobile Insurance Industry Using Ensemble Models

Abstract Fuzzy logic is a paradigm for an alternative design methodology, which can be applied in developing both linear and non-linear systems for embedded control. One of successful application that can use fuzzy control is liquid level control. This control system keeps developing from time to time to replace the ordinary system which applies mechanical functions in its control in order to improve the system reliability. In order to find the best design to stabilize the liquid level, the purpose of this thesis is to design a simulation system of fuzzy logic controller for water tank level control by using simulation GUI package which in MATLAB software. The water level was controlled by using three rules of membership function which then extended to five rules for verification purpose. This thesis is focus to the software part only, by doing some modification the design will be very useful for the system relates to liquid level control that widely used in industry nowadays. For a long time, the choice and definition of the parameters of PID controller are very difficult especially in nonlinear systems. There must be a bad effect if parameters wasn’t chosen nicely but the fuzzy controller can solve this problems because it’s based on the operator’s understanding of the behavior of the process (physical and dynamical response) instead of its detailed mathematical model to conclude the rules by the experts that match desired behavior. In this thesis, the water tank system was tested by both fuzzy and PID controller to analyze the control effect and comparing them with each other. As a result of comparing, fuzzy control is superior to PID control, which indicated that the fuzzy logic controller reduced the maximum overshoot and significantly reduced the steady state error and settling time as compared with PID controller.